T-705-Derived Prodrugs Show High Antiviral Efficacies against a Broad Range of Influenza A Viruses with Synergistic Effects When Combined with Oseltamivir

Emerging influenza A viruses (IAV) bear the potential to cause pandemics with unpredictable consequences for global human health. In particular, the WHO has declared avian H5 and H7 subtypes as high-risk candidates, and continuous surveillance of these viruses as well as the development of novel, broadly acting antivirals, are key for pandemic preparedness. In this study, we sought to design T-705 (Favipiravir) related inhibitors that target the RNA-dependent RNA polymerase and evaluate their antiviral efficacies against a broad range of IAVs. Therefore, we synthesized a library of derivatives of T-705 ribonucleoside analogues (called T-1106 pronucleotides) and tested their ability to inhibit both seasonal and highly pathogenic avian influenza viruses in vitro. We further showed that diphosphate (DP) prodrugs of T-1106 are potent inhibitors of H1N1, H3N2, H5N1, and H7N9 IAV replication. Importantly, in comparison to T-705, these DP derivatives achieved 5- to 10-fold higher antiviral activity and were non-cytotoxic at the therapeutically active concentrations. Moreover, our lead DP prodrug candidate showed drug synergy with the neuraminidase inhibitor oseltamivir, thus opening up another avenue for combinational antiviral therapy against IAV infections. Our findings may serve as a basis for further pre-clinical development of T-1106 prodrugs as an effective countermeasure against emerging IAVs with pandemic potential.


Introduction
Influenza A viruses (IAV) remain one of the major causative viral agents of acute respiratory illness and associated deaths worldwide. IAV can give rise to annual epidemics and occasional pandemics. According to estimations by the WHO, seasonal influenza epidemics may cause approximately 1 billion symptomatic infections worldwide, including 3-5 million severe cases and 290,000-600,000 influenza-related respiratory deaths every year [1]. Upon animal-to-man transmission, new IAV strains might emerge to which no or only limited immunity in a naïve human population exists. Further adaptation of these viruses to humans, allowing sustained human-to-human transmission, can lead to a pandemic of unpredictable magnitude and consequences for global health. The 3-Hydroxypyrazine-2-carboxylic acid 6. 3-Amino-2-pyrazinecarboxylic acid 5 (20.0 g, 144 mmol) was suspended in 100 mL conc. H 2 SO 4 . At 0 • C, sodium nitrite (14.9 g, 216 mmol) was added portion-wise. The suspension was afterward stirred at room temperature for two hours until a clear solution was formed. The reaction was quenched by purring the reaction mixture in ice water (500 mL). The precipitated solid was collected by vacuum filtration and washed with ice-cold water, and redissolved in a 5%-sodium carbonate solution. 3-Hydroxypyrazine-2-carboxylic acid 6 was precipitated by the addition of 1 M HCl (300 mL), collected by vacuum filtration, washed with ice cold water, and dried under vacuum to give 16.0 g of 6 (114 mmol, 77%) as a brownish solid and was used without further purification. 1 H-NMR (300 MHz, DMSO-d 6 ): δ [ppm] = 7.80 (d, 3 J H,H = 3.7 Hz, 1H, H-5), 7.65 (d, 3 J H,H = 3.7 Hz, 1H, H-6). Methyl-3-hydroxy-2-pyrazinecarboxylate 7. 6 (16.0 g, 114 mmol) was dissolved in MeOH (300 mL) and conc. H 2 SO 4 (2 mL) was added. After the reaction mixture was refluxed overnight all volatiles were removed under reduced pressure. Purification on silica gel (CH 2 Cl 2 /MeOH; 30:1) provided 14.0 g of 7 (90.8 mmol, 72%) as a yellow solid. 1 H-NMR (300 MHz, DMSO-d 6 ): δ [ppm] = 7.71 (d, 3 J H,H = 3.7 Hz, 1H, H-5), 7.47 (d, 3 J H,H = 3.7 Hz, 1H, H-6), 3.81 (s, 3H, CH 3 -Ester). (13) 0.2% BSA). Cells were washed once with sterile phosphate-buffered saline (PBS; Sigma-Aldrich, St. Louis, MO, USA) and infected for 30 min at 37 °C, followed by two additional PBS washing steps. Subsequently, 1 mL of infection medium (MEM, 1% Penicillin/Streptomycin, 0.2% BSA, and 1 µ g/mL TPCK-treated trypsin), supplemented with or without the test compounds, was added to each well, and the cells were further incubated for 24 h at 37 °C. At 24 h post-infection, cell culture supernatants were collected, and infectious virus titers were determined by plaque assay as described below.

Inhibitors
T-705 was purchased from ChemBlock (Hayward, CA, USA), and T-1105 from Accela ChemBio Inc. (San Diego, CA, USA). The nucleosides 2a, 2b, and 2c were synthesized based on previously reported protocols [29]. The derivatives 3a and 3b were synthesized according to published protocols at the chemistry department of the University of Hamburg, Hamburg, Germany [29]. Chemical synthesis pathways for those compounds, which are described for the first time in the present study, are described in detail below.

General Procedure 1: Synthesis of the Nucleosides 2a, 2b and 2c
For the synthesis of the nucleosides 2a, 2b, and 2c, a slightly modified protocol for the synthesis of 22 was used [29].
Step 1: Compound 7 (1.0 eq.) and ammonium sulphate (0.015 eq.) were suspended in hexamethyldisilazane (HMDS) (3 mL/mmol) and heated at 130 °C for 1 h until a clear solution was formed. After all volatiles were removed under reduced pressure, the residue was taken up in dry acetonitrile and CH2Cl. The appropriate ribose (0.7-0.8 eq.) followed by tin tetrachloride (1 M solution in CH2Cl2, 1.0 eq.) was added. The reaction mixture was stirred overnight at room temperature and then refluxed for a further 30 min. After the full conversion of the ribose could be observed, the reaction was quenched by the addition of saturated NaHCO3 solution. CH2Cl2 was added, and the organic layer was separated. The aqueous layer was extracted two times with CH2Cl2. The combined organic layers were dried over Na2SO4, and afterward all volatiles were removed under reduced pressure. Purification of the crude product on silica gel provided the protected nucleosides.
Step 2: The protected nucleosides from Step 1 were dissolved in ammonia (7 M in MeOH) and stirred overnight at room temperature. After all volatiles were removed under reduced pressure, the residue was dissolved in water. The aqueous layer was washed with CH2Cl2 three times and once with diethyl ether. After all volatiles were removed under reduced pressure, the crude product was purified on RP18 silica gel using automated flash chromatography (H2O/CH3CN; 100:0 to 0:100). T-1106 2a. According to general procedure 1, with 7 (1.19 g, 7.72 mmol), ammonium sulphate (15.3 mg, 0.12 mmol), and HMDS (23 mL). Then 1-O-acetyl-2,3,5-tri-O-benzoyl-beta-D-ribofuranose 14 (3.15 g, 6.24 mmol) and tin tetrachloride (7.72 mL, 7.72 mmol) in acetonitrile (55 mL) and CH2Cl2 (19 mL For the synthesis of the nucleosides 2a, 2b, and 2c, a slightly modified protocol for the synthesis of 22 was used [29]. Step 1: Compound 7 (1.0 eq.) and ammonium sulphate (0.015 eq.) were suspended in hexamethyldisilazane (HMDS) (3 mL/mmol) and heated at 130 • C for 1 h until a clear solution was formed. After all volatiles were removed under reduced pressure, the residue was taken up in dry acetonitrile and CH 2 Cl. The appropriate ribose (0.7-0.8 eq.) followed by tin tetrachloride (1 M solution in CH 2 Cl 2 , 1.0 eq.) was added. The reaction mixture was stirred overnight at room temperature and then refluxed for a further 30 min. After the full conversion of the ribose could be observed, the reaction was quenched by the addition of saturated NaHCO 3 solution. CH 2 Cl 2 was added, and the organic layer was separated. The aqueous layer was extracted two times with CH 2 Cl 2 . The combined organic layers were dried over Na 2 SO 4, and afterward all volatiles were removed under reduced pressure. Purification of the crude product on silica gel provided the protected nucleosides.

Synthesis of the Nucleosides 2d, 2e and 2f
First, 2d was synthesized with slight modifications as previously published. 7 (555 mg, 3.6 mmol), (2-acetoxyethoxy)methyl acetate (528 mg, 3.0 mmol) and by tin tetrachloride (1 M solution in CH2Cl2) (3.9 mL 3.6 mmol). Then, 581 mg of 2d (2.73 mmol, 91%) was received as a yellow oil. The analytical data are identical to those reported before [35]. 3 mmol) was added dropwise via a syringe, and the reaction mixture was stirred at 80 • C for 2 h. Water was added, and the solution was extracted twice with CH 2 Cl 2 . The combined organic layers were washed with water and brine, dried over Na 2 SO 4 , and concentrated under reduced pressure. The crude product was purified by column chromatography (CH 2 Cl 2 /MeOH; 50:1). Pure fraction of the Oand N-alkylated product were pooled and directly converted into the corresponding carboxamide using step 2 of general procedure 1. 247 mg of 2d (1.16 mmol, 39%) was received as an orange oil. 1  Next, 7 (693 mg, 4.5 mmol) and caesium carbonate (1.47 g, 4.5 mmol) in dry DMF were stirred at room temperature under nitrogen for 30 min. 4-Bromobutyl acetate 21 (433 µ L, 3 mmol) was added dropwise via a syringe, and the reaction mixture was stirred at 80 °C for 2 h. Water was added, and the solution was extracted twice with CH2Cl2. The combined organic layers were washed with water and brine, dried over Na2SO4, and concentrated under reduced pressure. The crude product was purified by column chromatography (CH2Cl2/MeOH; 50:1). Pure fraction of the O-and N-alkylated product were pooled and directly converted into the corresponding carboxamide using step 2 of general procedure 1. 247 mg of 2d (1.16 mmol, 39%) was received as an orange oil. 1

General Procedure 2: Nucleosides 18, 20 and 23
For the synthesis of the nucleosides 2a, 2b, and 2c, a slightly modified protocol for the synthesis of 22 was used [29]. To stir solution of protected nucleoside (1.0 eq.) in dry MeOH, sodium methanolate (5.4 M in MeOH, 0.1 eq.) was added dropwise. After the reaction mixture was stirred for 1 h at room temperature 1 M HCl was added until pH = 6. After all volatiles were removed under reduced pressure, the residue was dissolved in water and washed with CH2Cl2 four times and with diethyl ether one time. After all volatiles were removed under reduced pressure, the crude product was purified on RP18 silica gel using automated flash chromatography (H2O/CH3CN; 100:0 to 0:100).
T-1106 Methyl ester 22. According to general procedure 2, with 15 (7.  [29]. To stir solution of protected nucleoside (1.0 eq.) in dry MeOH, sodium methanolate (5.4 M in MeOH, 0.1 eq.) was added dropwise. After the reaction mixture was stirred for 1 h at room temperature 1 M HCl was added until pH = 6. After all volatiles were removed under reduced pressure, the residue was dissolved in water and washed with CH 2 Cl 2 four times and with diethyl ether one time. After all volatiles were removed under reduced pressure, the crude product was purified on RP18 silica gel using automated flash chromatography (H 2 O/CH 3 CN; 100:0 to 0:100).
T-1106 Methyl ester 22. According to general procedure 2, with 15 (7.  Step 1: To a stirred solution of nucleoside (1.0 eq.) and proton sponge (2.0 eq.) in trimethyl phosphate (TMP) was added dropwise POCl 3 at 0 • C. After the full conversion was observed by HPLC, the reaction was quenched by the addition of 1 M TEAB buffer. Afterward the reaction mixture was washed with CH 2 Cl 2 three times, and all volatiles were removed under reduced pressure. The crude product was purified on RP18 silica gel using automated flash chromatography (H 2 O/CH 3 CN; 100:0 to 0:100).

General Procedure 4: Synthesis of T-1106-TriPPPros 4a-4f
Step 1: To a stirred solution of H-phosphonate (1.0 eq.) in dry CH 3 CN (10 mL) Nchlorosuccinimide (2.5 eq.) was added. After the reaction mixture was stirred overnight at 50 • C, tetrabutylammonium phosphate monobasic solution (0.4 M in acetonitrile) (3.0 equiv.) was added dropwise. After the reaction mixture was stirred for one more hour, all volatiles were removed under reduced pressure, and the residue was dissolved in CH 2 Cl 2 and subsequently washed with 1 M ammonium acetate solution and cold water. Afterward, all volatiles were removed under reduced pressure to afford the corresponding pyrophosphate in quantitative yield.
Step 2: The corresponding pyrophosphate was dissolved in dry CH 3 CN (4 mL) and cooled to 0 • C. Afterward, a mixture of trifluoroacetic anhydride (5.0 equiv.) and Et 3 N (8.0 equiv.) in dry CH 3 CN (1.5 mL) was added dropwise. After the reaction mixture was stirred for 10 min at room temperature, all volatiles were removed under reduced pressure. Afterward, the residue was dissolved in dry CH 3 CN (6 mL) and cooled to 0 • C, and 1-methylimidazole was added. The reaction mixture was stirred for 10 min before the NMP (0.3 eq.) in dry DMF (2.5 mL) was added. The reaction mixture was stirred for 2 h at room temperature before all volatiles were removed under reduced pressure. The desired target molecules could finally be obtained after purification with RP18 silica gel using automated flash chromatography (H 2 O/CH 3 CN; 100:0 to 0:100) followed by ion-exchange to the ammonium form with Dowex 50WX8 cationexchange resin and a second chromatography step with RP18 silica gel using automated flash chromatography (H 2 O/CH 3 CN; 100:0 to 0:100).

General Procedure 4: Synthesis of T-1106-TriPPPros 4a-4f
Step 1: To a stirred solution of H-phosphonate (1.0 eq.) in dry CH3CN (10 mL) Nchlorosuccinimide (2.5 eq.) was added. After the reaction mixture was stirred overnight at 50 °C, tetrabutylammonium phosphate monobasic solution (0.4 M in acetonitrile) (3.0 equiv.) was added dropwise. After the reaction mixture was stirred for one more hour, all volatiles were removed under reduced pressure, and the residue was dissolved in CH2Cl2 and subsequently washed with 1 M ammonium acetate solution and cold water. Afterward, all volatiles were removed under reduced pressure to afford the corresponding pyrophosphate in quantitative yield.
Step 2: The corresponding pyrophosphate was dissolved in dry CH3CN (4 mL) and cooled to 0 °C. Afterward, a mixture of trifluoroacetic anhydride (5.0 equiv.) and Et3N (8.0 equiv.) in dry CH3CN (1.5 mL) was added dropwise. After the reaction mixture was stirred for 10 min at room temperature, all volatiles were removed under reduced pressure. Afterward, the residue was dissolved in dry CH3CN (6 mL) and cooled to 0 °C, and 1-methylimidazole was added. The reaction mixture was stirred for 10 min before the NMP (0.3 eq.) in dry DMF (2.5 mL) was added. The reaction mixture was stirred for 2 h at room temperature before all volatiles were removed under reduced pressure. The desired target molecules could finally be obtained after purification with RP18 silica gel using automated flash chromatography (H2O/CH3CN; 100:0 to 0:100) followed by ion-exchange to the ammonium form with Dowex 50WX8 cation-exchange resin and a second chromatography step with RP18 silica gel using automated flash chromatography (H2O/CH3CN; 100:0 to 0:100).

Virus Titration by Plaque Assay
Titration of infectious virus particles was carried out on MDCKII cells in 12-well tissue culture plates. Infectious cell culture supernatants were serially diluted 10-fold in PBS. Cells were washed once with PBS, and 150 µL of the appropriated dilution was added to the respective well. Cells were incubated for 30 min at 37 • C and subsequently covered with a 1:1 mixture of 2.5% Avicel (in double-distilled water) and 2× MEM medium (Fisher Scientific, Hampton, NH, USA), supplemented with 2% Penicillin/Streptomycin, 2% Lglutamine, 0.4% BSA and 1 µg/mL TPCK-treated trypsin. After incubation for 72 h at 37 • C, cells were washed once with PBS and then fixed with 4% paraformaldehyde solution for 30 min at 4 • C. Staining of H1N1 IAV plaques was performed with a primary influenza A nucleoprotein specific antibody (Abcam, Cambridge, UK) and a secondary HRP-conjugated antibody (Southern Biotech, Birmingham, AL, USA). Plaques were visualized by adding an HRP substrate (Seracare, Milford, MA, USA). H3N2, H5N1, and H7N9 IAV plaques were stained by counterstaining with crystal violet solution. Infectious virus titers were determined as plaque-forming units per ml (pfu/mL).

Cell Viability Assay
Measurement of cell viability upon compound treatment was performed using the CellTiter 96 ® Non-Radioactive Cell Proliferation Assay (MTT) (Promega, Madison, WI, USA) according to the manufacturer's instructions. In short, 1 × 10 4 MDCKII cells per well were seeded in 96-well plates. The cells were treated for 24 h with eight different inhibitor concentrations, diluted in an MDCKII cell culture medium. As a vehicle control, cells were treated with DMSO only. After incubation of the cells with the MTT-like substrate for 4 h, a stop solution was added, and the plates were further incubated at 37 • C overnight to allow solubilization of the blue crystals. Absorbance at 570 nm was measured on a Saphire2 plate reader (TECAN, Männedorf, Switzerland), and the cell viability was calculated in comparison to the DMSO control treatment.

Quantification and Statistical Analyses
All data generated in this study were analyzed and graphically visualized using Microsoft Excel (2016) and GraphPad PRISM Version 9.4.0 (RRID: SCR_002798). Statistically significant differences between vehicle controls and single-drug treatments at a defined concentration were determined using the Kruskal-Wallis test in combination with Dunn's multiple comparisons tests.
In order to investigate synergistic interactions of compound 3a and oseltamivir, the open-source web application SynergyFinder Plus was used [36]. Herein, four different reference models were employed to evaluate drug synergy: (i) the Bliss independence model (which expects that two drugs induce independent effects), (ii) the Zero Interaction Potency (ZIP) model (compares changes in the potency of the dose-response curves between individual drugs and their combinations), (iii) the highest single-agent (HSA) model (assumes that the expected effect of the combination treatment equals the maximum of the single drug responses), and (iv) the Loewe additivity (Loewe) model (which calculates the expected response in case both drugs were identical).

Chemical Synthesis of a T-1105/1106 Derivate Library Consisting of Nucleosides, Di-and Triphosphate Pronucleotides
In this study, we synthesized a library of T-705 derivates in order to identify new variants with increased antiviral activity against both seasonal and especially highly pathogenic avian influenza viruses (Figure 1). Herein, we pursued two different approaches: First, we developed DP (diphosphate) and TP (triphosphate) prodrugs of the T-1105 base, aiming to circumvent any cellular bottlenecks in nucleobase analogue activation (e.g., HGPRT and nucleotide kinase activity) (Figure 1). In order to achieve cell permeability of the negatively charged nucleotides, we introduced lipophilic masks at the terminal phosphate. After uptake into the cells, the release of the nucleotides is initiated by nucleophiles or esterase/lipase activity [29]. The released nucleoside analogue TPs can directly serve as a substrate for the viral RNA-dependent RNA polymerase (RdRP), whereas the DPs either act directly as substrates of the polymerase [29] or need to undergo further phosphorylation via cellular kinases (Figure 1).
Since the main mode of action of T-1106 is lethal mutagenesis, we aimed to modify the ribose part of T-1106 to enable chain termination as the predominant mechanism of action that might lead to enhanced antiviral activity. For that reason, we chose two different modifications, the 2 -C-methyl-and the 3 -deoxymodification. After a nucleoside analogue with a 2 -C-methylmodification is incorporated into the viral RNA strand, it prevents the RdRP from fluctuating into its closed conformation, which is necessary for the incorporation of the next incoming NTP (nucleotide triphosphate) [37]. The removal of the 3 -OH-group should lead to obligate chain termination due to the fact that the next incoming NTP has no functional group that it can bind. In addition to the modifications of the ribose part, we introduced a non-cleavable alkyl modification at the γ-phopshate group of the TP prodrugs. A previous study has shown that γ-alkyl-NTPs have a higher selectivity towards viral polymerases, which might reduce undesired cytotoxicity and side effects of the nucleoside analogues [38]. The T-1106-DiPPro-prodrugs 3a and 3b release the 5′-diphosphates instead, which are then metabolized into their 5′-triphosphates.

T-1106 Di-and Triphosphate Prodrugs Potently Inhibit Seasonal Influenza Virus Replication
The synthesis and antiviral activity of T-1105-derived DP and TP prodrugs against influenza A (A/X-31 (H3N2)) and influenza B (B/Ned/537/05) viruses were originally de-

T-1106 Di-and Triphosphate Prodrugs Potently Inhibit Seasonal Influenza Virus Replication
The synthesis and antiviral activity of T-1105-derived DP and TP prodrugs against influenza A (A/X-31 (H3N2)) and influenza B (B/Ned/537/05) viruses were originally described by Huchting et al. in 2018 [29]. Here, anti-influenza activity was measured by microscopic scoring of virus-induced cytopathic effects. In order to assess the potency of our T-705 derivate library to block the production of infectious virus progeny, we infected MDCKII cells with a seasonal H3N2 IAV isolate and determined virus titers in the presence of drug treatment. Reference compounds 1a (T-705) and 1b (T-1105) reduced virus replication by approximately 1.5 logs, in good agreement with previous reports (Figure 2a) [28,29]. T-1106 2a showed slightly higher activity compared to T-1105 1b. Surprisingly, any chemical modifications in the T-1106 nucleosides 2b-2f completely abrogated antiviral activity (Figure 2a). In contrast, at 50 µM, the DP prodrugs 3a and 3b reduced viral titers to the limit of detection, indicating high therapeutic efficacy. We further observed a pronounced difference in the antiviral activity of the TP prodrugs: while treatment with compound 4a showed a modest reduction in viral replication, albeit without reaching statistical significance, for TP compounds 4b-4f, no antiviral activity was detected (Figure 2a). Overall, these data confirm that T-1106-DP prodrugs 3a and 3b are highly active against H3N2 IAV replication in vitro, likely by overcoming bottlenecks in the intracellular conversion of T-705 1a and T-1105 1b to the metabolically active forms. In order to further characterize the antiviral properties of our lead DP (3a, 3b) and TP (4a) prodrug candidates, we determined their inhibitory concentration 50 (IC 50 ) against a more recent H1N1 IAV clinical isolate (Figure 2b-e). Compared to reference compound 1a (T-705), DP prodrug 3a showed a 10-fold increase in antiviral activity (IC 50 Table 1). Importantly, inhibition of viral replication at these concentrations was not caused by impaired cell viability upon drug treatment (Figure 2b-e; Table 1). It should be noted that as for H3N2 IAV (Figure 2a), maximum inhibition of H1N1 replication to the levels of detection limit was only achieved with the DP and TP prodrugs, but not with reference compound T-705 1a ( Figure S1a,d,g,j).   Table 1.
In summary, both DP and TP prodrugs are potent inhibitors of seasonal H1N1 and H3N2 IAV replication, with comparable or increased antiviral activity as the original and clinically licensed T-705 1a.
3.3. T-1106 Prodrugs 3a, 3b and 4a Are Highly Active against H5N1 and H7N9 HPAIV Next, we sought to analyze whether our lead DP and TP prodrug candidates (3a/3b and 4a, respectively) could also inhibit the replication of HPAIV (H5N1, H7N9) that have crossed the species barrier in the past and caused mortality rates of up to 40% in humans. Most importantly, these viruses are currently closely monitored by the WHO for the emergence of novel strains with pandemic potential. T-705 1a inhibited H5N1 replication with an IC 50 of 1.14 µM, while DP and TP prodrugs showed a 3-to 10-fold increase in antiviral activity (0.22, 0.12, and 0.31 µM IC 50 for 3a, 3b, and 4a, respectively) (Figure 3a-d; Table 1). Compared to 1a (0.79 µM IC 50 ), DP prodrug 3a was 10-fold more active against H7N9 (0.07 µM IC 50 ), while prodrugs 3b (0.62 µM IC 50 ) and 4a (2.02 µM IC 50 ) showed comparable or 2.5-fold reduced antiviral activity (Figure 3e-h; Table 1). As already mentioned above, the required concentrations to achieve inhibition of viral replication are non-cytotoxic to the cells, indicating that the observed antiviral activity is not a result of impaired cell viability. Of note, only DP prodrugs 3a and 3b reduced H5N1 and H7N9 virus titers to the limit of detection, in contrast to reference 1a and TP prodrug 4a ( Figures  S1b,c,e,f,h,k,l and S2).
These findings clearly demonstrate that metabolically pre-activated prodrugs of T-1106 are highly effective in inhibiting the replication of avian flu viruses with known pandemic potential. Importantly, compared to T-705 1a, DP prodrug compound 3a showed improved antiviral efficacy against various seasonal and avian influenza strains and thus might be an interesting candidate for further in vivo studies.   Table 1.

T-1106 Prodrug 3a Acts Synergistically with Oseltamivir against H5N1 HPAIV
Recent epidemic outbreaks of the highly pathogenic avian influenza virus H5N1 in birds, as well as sporadic infections of humans [3], have prompted us to investigate whether our lead prodrug candidate 3a could also act synergistically with existing therapeutics, such as the neuraminidase inhibitor oseltamivir. In order to define drug concentrations of 3a and oseltamivir suitable for studying synergistic interactions, we first determined the IC 50 of oseltamivir against H5N1 replication on MDCKII cells (0.008 µM IC 50 ; Figure S3a,b). Next, in addition to the IC 50 values, we calculated the inhibitory concentrations IC 1 , IC 10 , and IC 90 for DP prodrug 3a and oseltamivir, representing a 1%, 10%, or 90% inhibitory effect on viral replication, and used these concentrations for the combination treatment. We infected MDCKII cells with H5N1 at an MOI of 0.01 and therapeutically treated the virus-infected cells with a set of 16 different combinations of both compounds. In addition, we treated the cells with each drug alone (monotherapy). In terms of virus titer reduction, we observed a combinatory effect of the drug combinations that exceeded the efficacy of the respective monotherapies ( Figure S3c). We then used the open-source web application SynergyFinder Plus to evaluate synergistic interactions of DP prodrug 3a and oseltamivir. Herein, four different reference models (Bliss, ZIP, HSA, and Loewe) were used to calculate synergy scores that represent the strength of the observed interactions (a high synergy score indicates a strong synergistic effect). Landscape visualization, as well as the calculated synergy scores, revealed a strong synergistic effect of 3a when combined with oseltamivir, particularly at the two highest concentrations (IC 50 , IC 90 ) of both drugs (Figure 4). We further observed positive overall synergy scores independently of the reference model used, indicating, on average, a robust synergistic mode of action. Importantly, no antagonistic effects were detected ( Figure 4).
Overall, these data show that our lead DP prodrug compound 3a acts synergistically with oseltamivir against H5N1 avian influenza virus and thus could possibly be included in combination therapy in the context of pandemic preparedness.

Discussion
Current influenza treatment options licensed for clinical use are limited and include neuraminidase inhibitors (NAIs), such as oseltamivir, zanamivir, and peramivir; or drugs targeting the viral polymerase (viral polymerase inhibitors, VPIs), including baloxavir or the nucleoside analogue T-705 (Favipiravir) [4]. NAIs inhibit the function of the neuraminidase enzyme and thereby prevent the release of infectious virus particles from the cell [39]. Despite the emergence of drug-resistant viruses, particularly in immunocompromised patients [40][41][42][43], a rather short treatment window of the first two days after the onset of illness [44,45], as well as the uncertain benefit of treating severe influenza cases [6], NAIs are still recommended by various health-care institutions worldwide as the firstline antiviral therapy for influenza infections. VPIs, on the other hand, act on the viral

Discussion
Current influenza treatment options licensed for clinical use are limited and include neuraminidase inhibitors (NAIs), such as oseltamivir, zanamivir, and peramivir; or drugs targeting the viral polymerase (viral polymerase inhibitors, VPIs), including baloxavir or the nucleoside analogue T-705 (Favipiravir) [4]. NAIs inhibit the function of the neuraminidase enzyme and thereby prevent the release of infectious virus particles from the cell [39]. Despite the emergence of drug-resistant viruses, particularly in immunocompromised patients [40][41][42][43], a rather short treatment window of the first two days after the onset of illness [44,45], as well as the uncertain benefit of treating severe influenza cases [6], NAIs are still recommended by various health-care institutions worldwide as the first-line antiviral therapy for influenza infections. VPIs, on the other hand, act on the viral RdRP by various means, ultimately shutting down viral replication and blocking the production of virus progeny. Baloxavir, the only VPI approved in many countries for clinical use, inhibits the RdRP subunit PA (polymerase acidic protein) of various IAV subtypes, including H5N1 and H7N9 HPAIV, with pico-to nanomolar activity [46]. However, shortly after licensing, amino acid substitutions at position 38 in the PA protein (I38T/M/F) that decreased susceptibility to the drug were repeatedly observed in both H1N1 and H3N2 seasonal influenza viruses, suggesting a low barrier of resistance [8,9,47]. Favipiravir, alias T-705, is currently only licensed in Japan to treat IAV resistant to any other class of drug as part of pandemic preparedness. The mode of action was shown to be lethal mutagenesis and/or chain termination after consecutive incorporation, leading to the extinction of infectious virus particles [22,30]. T-705 needs to be converted into the active form by cellular enzymes; hence, this bottleneck in T-705 activation may lead to poor therapeutic efficacy, even at high treatment doses [26]. The therapeutic efficacy of T-705 is further decreased by its chemical instability once it is metabolized into the riboside form [27]. In this study, we aimed to overcome these limitations by chemically synthesizing nucleosides, di-and triphosphate prodrugs of the stabilized T-705 analogue T-1105 (termed T-1106 pronucleotides) that are able to release the active antiviral compounds directly into the infected cell. This was achieved by applying the Di-and TriPPPro-concept [31]. In previous studies, the chemical stability of the prodrugs was confirmed in either phosphate buffer at neutral pH or when incubated with cell extracts or blood serum [29,48]. Additionally, γmodified-TriPPPro-prodrugs were synthesized to achieve a higher selectivity toward viral polymerases. Furthermore, we introduced various chemical modifications (e.g., removal of the 3 -hydroxyl group or addition of a methyl group at the 2 -position of the sugar molecule) to enhance antiviral activity or to decrease the teratogenic/embryotoxic potential by deferring from lethal mutagenesis as the primary mechanism of action. Surprisingly, the majority of the introduced alterations abrogated antiviral activity against a human H3N2 IAV isolate, suggesting that they are no longer tolerated by the RdRP as alternative substrates for viral replication. In fact, we only observed the antiviral activity of T-705 (1a) or its de-fluoro analogue T-1105 (1b), the ribosylated form of T-1105 (T-1106; 2a), or three different DP and TP prodrugs (3a, 3b, and 4a). These prodrugs differ only in the length and size of the lipophilic mask but without any further chemical modifications at the sugar moiety. Improved antiviral efficacy of the DP and TP prodrugs against another H3N2 isolate was also shown by Huchting et al. in 2018 [29]. Furthermore, in this particular study, it was also demonstrated that bypassing cellular T-705 activation pathways is a viable strategy to retain T-705 activity even in the absence of key cellular enzymes involved in the T-705 metabolization [29]. Extending on these findings, here, we also show that DP and TP compounds 3a, 3b, and 4a potently block the replication of a more recent H1N1 IAV strain that was isolated during the 2009 swine flu pandemic. Most importantly, in the present study, we aimed to identify T-705/T-1105 derivates that possess broad-range antiviral activity against various IAV subtypes, including H5N1 and H7N9 HPAIV, which according to the WHO, bear pandemic potential. The efficacy of T-705 against these highly pathogenic viruses has already been demonstrated in vitro and in vivo [49][50][51]. Here, we show that DP and TP prodrugs 3a, 3b, and 4a are highly efficient in inhibiting the replication of both H5N1 and H7N9 HPAIV at low micromolar concentrations. However, we observed a clear impact of the number of phosphate groups attached to the sugar moiety: both DP prodrugs 3a and 3b showed 5-to 10-fold increased activity against all viruses tested, compared to the TP compound 4a and the original T-705 1a drug, which is consistent with previous reports [29]. The reduced efficacy of 4a might be best explained by impaired uptake into the cells due to the additional negatively charged phosphate atom; however, this is yet to be confirmed in future studies. Of note, while T-705 did not affect cell viability at high treatment doses, prodrugs 3a, 3b, and 4a caused considerable cell cytotoxicity starting at 30 µM, again in good agreement with the published literature [29]. Nevertheless, a reduction in viral titers to the limit of detection was observed at concentrations where no major adverse effects on cell viability could be detected. Future work needs to further address potential cytotoxic properties and side effects of the DP and TP prodrugs both in vitro and in vivo.
Prolonged treatment with any therapeutic as well as sub-optimal treatment doses, particularly in situations of compromised immune status such as immunosuppression or pregnancy, may lead to the in-host emergence of viral quasi-species with altered susceptibility to the drug used. This has been repeatedly seen in recent years for both NAIs and VPIs [52,53]. In this regard, it is important to achieve maximal virus inhibition at a safe dose to reduce the likelihood of resistant escape variants evolving under drug-selection pressure. While IC 50 values are often used to compare the efficacy of a given set of drugs, they do not provide information on residual virus replication in the presence of the treatment. Here, we show that only treatment with the DP and TP prodrugs can reduce both seasonal and HPAIV virus titers to the limit of detection, in sharp contrast to T-705. These findings suggest that T-1106-based prodrugs, which can act independently of cellular metabolic bottlenecks, might be superior to T-705 in terms of preventing the emergence of resistant viruses. It should be noted, however, that even though T-705 resistance could be achieved in cell culture [54], so far, no resistance-conferring amino acid substitutions have been detected in animal models [55] or in clinical trials [56]. Nevertheless, due to the restricted usage of T-705 in humans, it is still unclear whether prolonged and widespread use of the drug might lead to the emergence of resistant influenza viruses.
The ongoing COVID-19 pandemic has revived the debate on whether combination therapy using antiviral agents that differ in their mechanism of action is of clinical interest for treating (severe) influenza infections [57,58]. Combining different drugs may have various benefits over classical monotherapy, such as potentiating the overall antiviral effect, increasing the barrier to viral resistance (particularly within immunocompromised patients), and reducing the effective doses for each individual drug, thereby avoiding the possibility of adverse side effects [57]. Numerous previous reports have already demonstrated synergistic interactions of T-705 with other licensed anti-influenza drugs, such as oseltamivir or baloxavir, both in vitro as well as in animal models and in humans [56,[59][60][61][62]. Recent epidemic outbreaks of H5N1 HPAIV in wild birds and sporadic spillovers to mammals, including humans, particularly in the US and in South America [3], have prompted us to investigate potential drug synergies of our lead T-1106 DP prodrug candidate 3a with oseltamivir against H5N1 HPAIV. Here, we demonstrate that combination treatment effectively reduces H5N1 viral replication to a level that clearly exceeds the effect of the respective monotherapies, confirming a synergistic mode of action. Even though our findings provide the first evidence that T-1106-derived prodrugs may be suited for combinational therapy with oseltamivir, we acknowledge that further in-depth studies using other IAV subtypes as well as other NAIs are required to confirm broad synergistic activity. In addition, possible negative drug-drug interaction effects of the combination treatment need to be evaluated in an appropriate animal model, along with confirming an additive benefit of the drug combination over the respective monotherapies in vivo. Future studies should also address the risk of acquiring resistance against either drug when used as a combination therapy.
In Japan, T-705 use in humans is largely restricted to novel IAVs with resistance to existing drugs. This is mainly due to some early reports on potential teratogenic and embryotoxic properties [25]. Recent studies suggest that Favipiravir might be considerably safer also during pregnancy; however, due to the small number of participants in each study, it is not possible to draw any definite conclusions at this stage [63,64]. We are currently testing both T-705 and our lead prodrug candidates in an allogenic mouse pregnancy model [65] to shed further light on the safety profile of T-705 and its derivates in vivo.
To our knowledge, this is the first study that demonstrates the potent antiviral efficacy of T-705/T-1105-derived prodrugs not only against seasonal but also highly pathogenic avian influenza viruses. Thus, the findings of this study might provide a basis for antiviral therapies against HPAIV, which pose a high pandemic threat.